Prostaglandins Maintain Renal Vasodilation and Hyperfiltration During Chronic Nitric Oxide Synthase Blockade in Conscious Pregnant Rats
Rats demonstrate renal vasodilation and hyperfiltration in pregnancy. Because both NO and cGMP biosynthesis are increased in gravid rats and because acute administration of NO synthase inhibitors abrogates renal vasodilation and hyperfiltration, NO most likely mediates the renal circulatory changes of gestation. In the present study, we tested the effect of chronic inhibition of NO synthase on effective renal plasma flow (ERPF) and glomerular filtration rate (GFR) in chronically instrumented, conscious, gravid rats. Because gestation is a relatively long-term condition, we postulated that chronic withdrawal of NO would result in sustained inhibition of renal vasodilation and hyperfiltration. Contrary to our hypothesis, the renal circulatory changes of pregnancy were maintained during chronic blockade of NO synthase. That is, subcutaneous administration of 10 μg/min Nω-nitro-l-arginine methyl ester (NAME) for 48 hours did not significantly reduce GFR in either virgin or pregnant rats; thus, hyperfiltration persisted in the latter despite chronic NO synthase blockade. In contrast, ERPF was reduced and effective renal vascular resistance (ERVR) increased in both groups of rats during NAME administration but in a parallel fashion, such that renal vasodilation persisted in the gravid animals despite chronic inhibition of NO synthase. However, with superimposition of acute prostaglandin synthesis inhibition (meclofenamate, 10 mg/kg IV), renal vasodilation and hyperfiltration were abolished; ie, the combined treatments of chronic NO synthase blockade and acute prostaglandin synthesis inhibition led to the equalization of GFR, ERPF, and ERVR in conscious virgin and pregnant rats. Inhibition of prostaglandin synthesis alone had little affect on the renal circulation, as previously reported. In summary, prostaglandins are recruited to maintain renal vasodilation and hyperfiltration during chronic NO synthase blockade in conscious pregnant rats.
Normal human gestation is characterized by significant vasodilation of the maternal circulation. The renal vascular bed participates in this profound vasodilatory response to pregnancy. During the late first or early second trimester, renal blood flow and GFR peak at 40% to 80% above prepregnant values (reviewed in Reference 1). Additionally, the increase of renal blood flow is in the face of a 5 to 10 mm Hg decline in diastolic blood pressure,2 underscoring the marked decrease in renal vascular resistance that occurs during human gestation. Comparable to gravid women, the chronically instrumented conscious pregnant rat undergoes both systemic and renal vasodilation. Rat gestation is accompanied by a significant decrease in renal vascular resistance resulting in a 20% to 40% increase in GFR and renal blood flow above prepregnant values.3
Ultimately, renal vasodilation in pregnancy is likely to be under endocrinologic control. Although vasodilatory prostaglandins are unlikely mediators of the renal hemodynamic changes in unstressed chronically instrumented rats during pregnancy,4 5 6 NO and cGMP appear to be intimately involved.7 8 9 10 In particular, earlier work demonstrated that acute blockade of NO synthase inhibited renal vasodilation and hyperfiltration in chronically instrumented conscious pregnant rats.10 In the present study, we investigated the effect of chronic inhibition of NO synthase on renal hemodyamics and the GFR in conscious gravid rats. Because pregnancy is a relatively long-term condition, we reasoned that chronic withdrawal of NO would result in sustained blockade of gestational renal vasodilation and hyperfiltration.
Materials and Methods
Female Long-Evans rats, aged 12 to 14 weeks, were purchased from Harlan Sprague Dawley (Madison, Wis) and fed Harlan Teklad rat chow (No. 591355). The rats were habituated to a specially designed experimental cage from 1 to 3 hours per day for a minimum of 5 days before surgical preparation. This cage allowed for normal grooming of the face and front paws while preventing the rat from turning around. Thus, accurate timed urine collections were possible via the chronically implanted bladder catheter (described below). Rats that failed to acclimate to the restraining cage were eliminated from the study (<2% of all animals). Details of the surgical procedure have been previously described.3 4
Briefly, with rats under ketamine (6 mg/100 g body wt) and pentobarbital sodium (2.1 mg/100 g body wt) anesthesia and by use of aseptic technique, Tygon catheters (internal diameter, 0.015 inch; outer diameter, 0.030 inch; Norton) were implanted in the abdominal aorta and inferior vena cava via the femoral artery and vein, respectively. The tips of these catheters lay below the level of the renal vessels. The catheters were then routed subcutaneously with a trochar, exteriorized between the scapulae, filled with a 1:1 mixture of 50% dextrose and 1000 U/mL heparin, and plugged with stainless steel pins. A cannula of silastic-covered stainless steel was then sewn into the urinary bladder with a purse-string suture and exteriorized through the ventral abdominal wall. This bladder cannula was stopped with a silastic-covered 18-gauge pin so that the rat voided normally through the urethra while in her home cage. For experiments, the bladder cannula pin was removed for urine collection. The rats were housed individually after completion of the surgery and allowed 10 to 14 days of recovery before mating. The vascular catheters were then allowed to slip under the skin in all animals. By random selection, half of the rats were housed with competent males; the others served as age-matched virgin controls. The presence of spermatozoa in the vaginal lavage marked day 0 of gestation (term pregnancy, 22 days). On gestational day 4 or 5, the catheters were exteriorized under light ether anesthesia, flushed with Ringer's solution, and refilled with the dextrose/heparin mixture. Measurement of baseline renal function was then recorded on gestational days 10 to 12, when renal hemodynamics and GFR are maximal in this species.3 These procedures were in accordance with our institutional guidelines.
Administration of NAME (7 Virgin and 7 Pregnant Rats)
For each experiment, baseline renal function was assessed concurrently in an age-matched virgin control and pregnant rat on gestational day 10, 11, or 12. The arterial catheter was connected to a Gould P23 ID pressure transducer (Statham Instruments) and a Gould pressure processor. Phasic and mean blood pressures and heart rate were displayed on three channels of a Gould TA11 chart recorder. This catheter was also used for collection of blood samples at the midpoint of each timed urine collection. The femoral venous catheter was used to administer a priming bolus, followed by constant infusion of inulin (5 g/50 mL, Cyprus Pharmaceutical Co) and p-aminohippurate (20% sodium hippurate, Merck Sharpe & Dohme) at a flow rate of 10 μL/min delivered by a Sage infusion pump (model 355, Orion Research). Finally, the obturator in the bladder cannula was removed and extended with a short piece of polyethylene tubing, and timed urine collections were made. This technique of urine collection proved to be reliable: after reaching steady state conditions, the urinary excretion rates of inulin and p-aminohippurate were 99±2% and 99±2% of their respective infusion rates for virgin rats (n=21) and 98±2% and 100±2% for the gravid rats (n=19). The renal clearance of inulin and p-aminohippurate thus provided an accurate measure of GFR and ERPF, respectively.
After the priming bolus and constant infusion of inulin (0.2 mL of 10% stock/100 g body wt and 0.5 mg·min−1·100 g body wt−1, respectively) and p-aminohippurate (0.2 mL of 1% PAH/100 g body wt and 0.1 mg·min−1·100 g body wt−1, respectively) were initiated, an equilibration period of 60 minutes was allowed for the compounds to reach steady state plasma concentration. Then, three 30-minute baseline urine collections were made with midpoint blood samples (200 μL each). After centrifugation of blood samples and separation of plasma from cells, the latter were resuspended in Ringer's solution and returned to the rat. The infusion was then discontinued, and both vascular catheters were flushed with Ringer's solution, filled with the dextrose/heparin mixture, and stoppered with stainless steel pins. Next, the bladder was rinsed with sterile water followed by a neomycin/polymyxin B irrigation mixture (1:10, Burroughs Wellcome Co) and plugged with the silastic-covered 18-gauge pin. The rat was returned to her home cage.
After 24 hours, both gravid and virgin control rats were lightly anesthetized with ether. An Alzet miniosmotic pump (Alza) containing NAME (Sigma Chemical Co) HCl was inserted subcutaneously into the back of the animal, and the incision was closed with staples. The miniosmotic pumps delivered NAME at a rate of 10 μg/min, a dosage that significantly inhibits NO synthesis.9 Forty-eight hours later, renal function was again assessed during the chronic inhibition of NO synthase.
Time-Control Studies (7 Virgin and 7 Pregnant Rats)
Rats implanted with miniosmotic pumps containing 0.9% saline (vehicle) for 48 hours served as time controls for the rats receiving NAME treatment. These rats were subjected to the same experimental protocol as the NAME-treated rats. They were similarly tested in pairs, one pregnant and one virgin, on the same day and with the same renal parameters being measured.
Administration of NAME and/or Meclofenamate (7 Virgin and 5 Pregnant Rats)
Pregnant and age-matched virgin control rats underwent training to the experimental cage, mating, and surgical preparation as described above. Baseline renal function was assessed in virgin control and pregnant rats concurrently. In order to inhibit prostaglandin synthesis,4 an infusion of meclofenamate (10 mg/kg body wt, Sigma) was administered over 10 minutes via the venous catheter, followed by a 30-minute equilibration. Three additional 30-minute renal clearances were performed. After 24 hours, the miniosmotic pumps containing NAME (10-μg/min infusion rate) were implanted subcutaneously in the back of each animal. After 48 hours of NAME administration, three 30-minute renal clearances were obtained, followed by an infusion of meclofenamate (10 mg/kg body wt) and three additional 30-minute renal clearances.
Inulin in plasma and urine was measured by the anthrone method,11 and p-aminohippurate was measured by the method of Bratten and Marshall as modified by Smith et al.12 Urine volume was determined gravimetrically. All renal clearance data have been expressed per whole animal. For assay of urinary NOx, dilutions of each urine sample were incubated for 1 hour with nitrate reductase, NADPH, and flavin adenine dinucleotide in a potassium phosphate buffer.13 Duplicate aliquots of each urine sample were then mixed with Griess reagent (1% sulfanilamide and 0.1% naphthylethylenediamine dihydrochloride) to form a stable azo dye, which was measured by spectrophotometry at 546 nm. Details of the NOx assay and assay validation have been previously published.9
In view of the recent report that nitro-containing l-arginine analogues interfere with the measurement of NOx,14 we tested NAME in our assay. NAME was substituted for nitrate in the standard curve (1.4 to 7.2 μmol/L) and failed to produce nitrite. Concentrations of NAME as high as 1.0 mmol/L did not affect the nitrate standard curve. Thus, nitrate reductase does not recognize NAME as a substrate: NAME does not interfere with the enzymatic reduction of nitrate to nitrite or the assay of NOx.
Urinary 6-keto-PGF1α was assessed by direct radioimmunoassay after dilution of samples 3-fold and 6-fold with assay buffer. The antibody was obtained from Perceptive Diagnostics and diluted 5500-fold for use as recommended by the manufacturer. The 3[H]ligand was purchased from New England Nuclear, and the standard was from Sigma. The standard curve ranged from 10 to 10 000 pg per tube, and the assay sensitivity was ≈50 pg per tube.
Preparation of Drugs
NAME HCl and meclofenamate were freshly prepared for each experiment. Inulin and p-aminohippurate were also prepared on the morning of the experiment using Ringer's solution as a diluent. Inulin, characteristically insoluble at room temperature, was prepared for infusion by heating a 15 mL aliquot in a boiling water bath for 3 minutes. When diluted and mixed with p-aminohippurate, it remained in solution throughout the experiment.
Data are presented as mean±SEM. Differences in baseline parameters between virgin and midterm pregnant rats (Table 1⇓) were analyzed by unpaired Student's t tests. The urinary excretion rates of 6-keto-PGF1α and NOx and changes in renal parameters during combined blockade of NO and prostaglandin synthesis were analyzed by the Mann-Whitney U test. All other data were analyzed by two-factor repeated measures ANOVA. If significant main effects were observed, then individual group means were contrasted by the Tukey test or orthogonal contrasts. A value of P<.05 was taken to be significant.15
Table 1⇑ summarizes the baseline values for all virgin and midterm pregnant rats. Both GFR and ERPF were significantly increased by ≈25% in the gravid rats compared with virgin control rats. ERVR was reduced by a similar degree in the pregnant rats.
Renal function was stable over 48 hours of vehicle administration in virgin and gravid rats, as shown in the time-control studies (Table 2⇓). The effect of NO synthase inhibition alone on renal function using the arginine analogue NAME is depicted in Table 3⇓. Before infusion of the NO synthase inhibitor, GFR and ERPF were significantly greater and ERVR was less in the gravid rats compared with the virgin rats. During 48-hour inhibition of NO synthase, MAP and ERVR were significantly increased in both gravid and virgin rats. The GFR was not greatly affected by chronic NAME infusion in either the gravid or virgin rats, but ERPF was significantly decreased in both groups of animals. Inspection of the changes in MAP, ERPF, and ERVR from baseline shows that the pregnant and virgin rats were comparably affected by the NAME infusion. Thus, GFR and ERPF remained higher and ERVR remained lower in the gravid rats compared with virgin control rats during chronic inhibition of NO synthase.
At baseline, the urinary NOx excretion was significantly higher in the gravid animals compared with the virgin controls (Table⇑s 2 and 3, and as previously reported9 ). During the infusion of NAME, urinary NOx excretion was reduced to the same level in gravid and virgin rats (Table 3⇑). Time-control experiments over a 48-hour period of vehicle administration revealed no significant change in urinary NOx excretion for either the virgin or pregnant rats (Table 2⇑).
The influence of acute prostaglandin synthesis inhibition, chronic NO synthase inhibition, and the combination thereof on renal function is depicted in Figs 1 through 4⇓⇓⇓⇓. MAP was unaffected in either gravid or virgin animals by meclofenamate alone (Fig 1⇓). After a 48-hour infusion of NAME, MAP increased significantly in gravid and virgin animals in a parallel fashion (P<.001), corroborating the initial study (Table 3⇑). Subsequent administration of meclofenamate did not further affect MAP.
Baseline GFR for gravid rats was significantly higher than in the virgin controls (Fig 2⇑). For both groups of rats, GFR was not significantly affected by either acute meclofenamate or chronic NAME infusion. However, GFR was virtually equalized in the two groups of rats after acute administration of meclofenamate superimposed on the chronic infusion of NAME. The absolute decline in GFR for pregnant rats exceeded that of the virgin control rats (−700±251 versus −363±163 μL/min, respectively), although statistical significance was not reached.
ERPF was also significantly greater in the pregnant rats compared with virgin controls (Fig 3⇑). ERPF showed a minor decline after acute infusion of meclofenamate in both groups of rats. In contrast, a major decline was observed after 48-hour inhibition of NO synthase (P<.005). However, ERPF remained significantly greater in the gravid rats until the acute administration of meclofenamate. That is, the combined treatments of chronically administered NAME and acutely administered meclofenamate produced a convergence of ERPF in gravid and virgin rats. The fall in ERPF for pregnant rats greatly surpassed that of the virgin control rats (−3283±30 versus −375±468 μL/min, respectively; P<.005).
Fig 4⇑ depicts the changes in ERVR. Midterm pregnant rats showed a significantly lower ERVR than virgin control rats. After prostaglandin synthesis inhibition, both groups of rats demonstrated small insignificant increases in ERVR. A large increase of ERVR was observed in both gravid and virgin rats in response to the 48-hour infusion of NAME, again corroborating the initial study (Table 3⇑). Part of this increase in ERVR results from autoregulatory adjustments in response to the rise in MAP. However, the gravid rats remained relatively vasodilated compared with the virgin controls during chronic inhibition of NO synthase. Only with subsequent administration of meclofenamate did ERVR converge in the two groups of rats, mainly because of a more robust response by the gravid rats (+9.34±2.54 versus +3.50±1.46 mm Hg·mL−1·min−1, P<.05).
To substantiate inhibition of prostaglandin and NO synthesis in this series of experiments, 6-keto-PGF1α and NOx, respectively, were measured in the urine of both the midterm pregnant and virgin rats. The urinary excretion of 6-keto-PGF1α is portrayed in Table 4⇓. Comparable reductions after meclofenamate administration were observed for the two groups of rats. Interestingly, a 2- or 3-fold increase in urinary excretion of 6-keto-PGF1α was observed after 48-hour infusion of NAME. The urinary excretion of NOx was measured in 3 of the midterm pregnant and virgin rats. As observed previously (Table 3⇑), the urinary excretion of NOx was reduced overall by 74±10% after the 48-hour administration of NAME (from 12.5±2.2 to 4.3±1.3 nmol/min, P<.02).
Endothelium-derived relaxing factors have been proposed as mediators of the systemic and renal vasodilatory responses of pregnancy. The urinary excretion of several prostaglandins or their major metabolites increases during gestation, suggesting enhanced endogenous biosynthesis and a possible role for these factors in the vasodilatory responses (see Reference 6 and citations therein). However, a critical role for vasodilatory prostaglandins in the control of the circulation during pregnancy proved unlikely, insofar as systemic and renal vasodilation and hyperfiltration persisted despite prostaglandin synthesis inhibition.4 5 16 Consistent with this conclusion was the finding that renal tissues from gravid rats failed to show augmented production of various prostaglandins ex vivo.6 The endogenous biosynthesis of NO and of its second messenger, cGMP, also increases during gestation (References 7 through 9 and Tables 2 and 3⇑⇑ of the present study). In contrast to the prostaglandins, a critical role for NO in the control of the renal circulation during pregnancy was established, insofar as acute administration of NO synthase inhibitors abrogated renal vasodilation and hyperfiltration in chronically instrumented conscious pregnant rats.10
Because pregnancy is a relatively long-term physiological condition, our current objective was to investigate the impact of chronic inhibition of NO synthase on the renal circulation in conscious gravid rats. We postulated that chronic withdrawal of NO would produce sustained inhibition of the renal vasodilation and hyperfiltration of pregnancy. Contrary to our expectation, however, renal vasodilation and hyperfiltration persisted relative to the virgin control rats, despite the reduction in urinary NOx excretion to comparable levels in the two groups of animals (Table⇑s 2 and 3). Thus, we reasoned that in contrast to acute blockade of NO synthase, chronic blockade may permit sufficient time for induction of compensatory vasodilatory mechanism(s), eg, the prostaglandins.
Consistent with the earlier reports,4 5 prostaglandin synthesis inhibition alone had little, if any, impact on the renal circulation in conscious unstressed nonpregnant and pregnant rats (Figs 2 through 4⇑⇑⇑). Previous investigations demonstrated that the imposition of various modes of stress (eg, anesthesia, surgery, or volume contraction) renders the renal circulation dependent on vasodilatory prostaglandins (reviewed in Reference 17). Moreover, gravid rats subjected to the stress of anesthesia and surgery may become more reliant than nonpregnant rats on vasodilatory prostaglandins, in order to defend the low vascular resistance of pregnancy.4 In view of this precedent for recruitment of prostaglandins to maintain systemic and renal vasodilation as a compensatory mechanism, we tested whether prostaglandins compensate for NO withdrawal during chronic blockade of NO synthase. The results confirmed the supposition, showing that acute inhibition of prostaglandin biosynthesis superimposed on the chronic blockade of NO synthase further compromised renal hemodynamics, particularly in gravid rats. The combined treatment virtually abolished the renal vasodilation and hyperfiltration of pregnancy (Figs 2 through 4⇑⇑⇑).
MAP was comparably increased in the virgin and midterm pregnant rats by chronic NAME administration (Table 3⇑ and Fig 1⇑). This finding is consistent with our previous work10 using acute administration of NAME (and monomethyl-l-arginine) but inconsistent with the report by Molnar et al.18 These investigators observed a greater pressor response to chronic administration of NAME in late gravid rats. There are several possible differences in experimental design that could account for this apparent inconsistency, but perhaps most noteworthy is the difference in gestational age when the rats were studied, ie, midterm versus late pregnancy. In the present work, superimposition of acute meclofenamate infusion on chronic NO synthase inhibition did not further raise MAP in either the virgin or midterm pregnant rats (Fig 1⇑), despite the equalization of ERVR in the two groups of animals, as described above. Without concurrent measurement of cardiac output, however, we cannot reliably comment about the status of total peripheral vascular resistance in the present study.
The mechanism by which chronic inhibition of NO synthase leads to the recruitment of prostaglandins that, in turn, maintain the renal circulatory changes of pregnancy is unknown. NO can either stimulate19 20 or inhibit20 eicosanoid production by various types of cells in culture. Interestingly, the urinary excretion of 6-keto-PGF1α (a major metabolite of prostacyclin) was increased in both the virgin and pregnant rats after the 48-hour treatment with NAME. It is possible that the increases in blood pressure produced by chronic inhibition of NO synthesis (Fig 1⇑) stimulated endothelial production of prostacyclin by a mechanical stress mechanism.21 In view of the considerable variability in the urinary excretion of 6-keto-PGF1α, we were unable to discern if the gravid rats excreted more of the prostacyclin metabolite than the virgin rats during the chronic blockade of NO synthase. If so, then possibly chronic inhibition of NO synthase stimulates prostaglandin production to a greater degree in the pregnant condition at vascular sites relevant to the control of renal hemodynamics, which then assume an important compensatory role in maintaining renal vasodilation and hyperfiltration. However, any differences in prostaglandin production cannot be easily attributed to the blood pressure increments induced by NAME, because these were comparable in virgin and pregnant rats (Fig 1⇑). Another consideration is that the renal vasculature may become more sensitive to the vasodilatory action of prostaglandins in the gravid rats subjected to chronic NO synthase blockade. Clearly, further work is needed to determine the mechanism of the prostaglandin compensation observed in the present study.
This investigation highlights the versatile nature of the pregnant condition, which is replete with overlapping and compensatory mechanisms or backup systems that are recruited in the event of failure of the primary system. In this case, NO is the primary mediator of renal vasodilation and hyperfiltration in pregnancy,10 whereas the prostaglandins serve in a fail-safe capacity. It is conceivable that if both of these vasodilatory pathways were chronically inhibited, then yet another vasodilatory substance, perhaps an endothelium-derived relaxing factor, would be recruited. The present study may have implications for the hypertensive disease of pregnancy, preeclampsia, in which renal hemodynamics and GFR22 as well as NO23 and prostacyclin24 biosynthesis may be compromised. That is, any further reduction in prostaglandin production by injudicious administration of nonsteroidal anti-inflammatory drugs may further compromise renal function in the disease.
Selected Abbreviations and Acronyms
|ERPF||=||effective renal plasma flow|
|ERVR||=||effective renal vascular resistance|
|GFR||=||glomerular filtration rate|
|MAP||=||mean arterial pressure|
|NAME||=||Nω-nitro-l-arginine methyl ester|
|NOx||=||nitrate and nitrite|
This study was supported by grants from the National Institutes of Health: RO1 HD-30325, RCDA KO4-01098, and Minority Institutional Research Training grant 5-T32-HL-07736. We thank Monique Mosher, Fran Lutka, and Leslie Minich for technical support in the assays of urinary NO and prostacyclin metabolites. The expert secretarial support of Carmen Redman is also gratefully acknowledged. Portions of this work served to partly fulfill the requirements of the PhD degree for L.A. Danielson.
Previously published in part in abstract form (J Am Soc Nephrol. 1996;7:1579).
- Received May 23, 1996.
- Accepted September 13, 1996.
Conrad KP. Renal changes in pregnancy. Urol Ann. 1992;6:313-340.
Conrad KP, Colpoys MC. Evidence against the hypothesis that prostaglandins are vasodepressor agents of pregnancy. J Clin Invest. 1986;77:236-245.
Baylis C. Renal effects of cyclooxygenase inhibition in the pregnant rat. Am J Physiol. 1987;253:F158-F163.
Conrad KP, Dunn MJ. Renal synthesis and urinary excretion of eicosanoids during pregnancy in rats. Am J Physiol. 1987;253:F1197-F1205.
Conrad KP, Vernier KA. Plasma level, urinary excretion, and metabolic production of cGMP during gestation in rats. Am J Physiol. 1989;257:R847-R853.
Conrad KP, Joffe GM, Kruszyna H, Kruszyna R, Rochelle LG, Smith RP, Chavez JE, Mosher MD. Identification of increased nitric oxide biosynthesis during pregnancy in rats. FASEB J. 1993;7:566-571.
Danielson LA, Conrad KP. Acute blockade of nitric oxide synthase inhibits renal vasodilation and hyperfiltration during pregnancy in chronically instrumented conscious rats. J Clin Invest. 1995;96:482-490.
Smith HW, Finkelstein N, Aliminosa L, Crawford B, Graber M. The renal clearances of substituted hippuric acid derivatives and other aromatic acids in dog and man. J Clin Invest. 1945;24:388-404.
Zar JH. Biostatistical Analysis. Englewood Cliffs, NJ: Prentice Hall; 1984.
Conrad KP, Dunn MJ. Renal prostaglandins and other eicosanoids. In: Windhager EE, ed. Handbook of Physiology, Section 8: Renal Physiology, Volume II. New York, NY: Oxford University Press; 1991.
Davidge ST, Baker PN, McLaughlin MK, Roberts JM. Nitric oxide produced by endothelial cells increases production of eicosanoids through activation of prostaglandin H synthase. Circ Res. 1995;77:274-283.
Davies PF, Tripathi SC. Mechanical stress mechanisms and the cell: an endothelial paradigm. Circ Res. 1993;72:239-245.
Conrad KP, Mosher MD. Nitric oxide biosynthesis in normal and preeclamptic pregnancy: a preliminary report. J Am Soc Nephrol. 1995;6:657. Abstract.
Fitzgerald DJ, Entman SS, Mulloy K, FitzGerald GA. Decreased prostacyclin biosynthesis preceding the clinical manifestation of pregnancy-induced hypertension. Circulation. 1987;75:956-963.